Lape_LowR pdf - University of Washington



Lape_LowR pdf - University of Washington
Archaeol. Oceania 43 (2008) 11–21
Fortification as a human response to late Holocene climate change in East Timor
Keywords: Climate change, ENSO, fortifications, warfare, East Timor
We present evidence for a significant shift in human
landscape use in post 1000 AD East Timor towards fortified
and defensively-oriented settlement sites. We propose a
model of agents selecting to invest in fortification building
that is based on the spatial and temporal variation in the
availability of rainfall-dependent resources. These resources
may have been significantly impacted by climatic events
associated with ENSO variation, and we discuss spatial and
temporal correlation with ENSO warm phase frequency and
dates of initial fortification building.
Hundreds of stone walled structures are visible on hilltops
and cliff edges in the contemporary landscape of north
coastal East Timor (Figure 1). These structures are remembered by local inhabitants as places occupied in the past
during a time of internecine warfare, mostly before living
memory. Recent archaeological work supports the
interpretation that these structures served as fortified
settlements, and that people shifted suddenly to this
settlement pattern only after 1000 AD, despite over 35,000
years of human occupation of Timor (e.g. Lape 2006). In
this paper, we propose a model based on rainfall dependent
resource distribution across time and space that could
explain this sudden shift in settlement pattern, and then
apply paleoenvironmental and recently collected
archaeological data to test this model. Of particular interest
is the history of the El Niño Southern Oscillation (ENSO),
which may have had considerable impact on East Timor
during the past 1000 years. In recent decades, El Niño
caused drought in much of Island Southeast Asia and the
Western Pacific. Although there are not yet any
paleoenvironmental data from East Timor itself, a variety of
proxy data from regions with demonstrated teleconnections
to East Timor’s climate indicate that El Niño events were
much more frequent in the past than in the 20th century.
Below we will review the culture history of East Timor,
introduce our model, test it with archaeological and
paleoenvironmental data and discuss the results.
PL: Department of Anthropology, University of Washington,
Seattle, Washington, 98195 USA. [email protected];
CC: Institute of History and Philology, Academia Sinica,
Nankang, Taipei 115, Taiwan, R.O.C. [email protected]
Culture history of East Timor in a regional context
Prior to its separation and subsequent independence from
Indonesia in 1999, East Timor had seen relatively little
attention from archaeologists. Most work published to date
has been conducted on an uplifted limestone plateau in the
easternmost Lautem district, where dates of human
occupation as early as 35,000 BP have been identified in
several solution cave sites (O’Connor 2003; O’Connor and
Veth 2005; O’Connor, et al. 2002; Spriggs, et al. 2003). The
rich rock art of this region has also been subject to recent
investigation (Lape, et al. in press; O’Connor 2003;
O’Connor and Oliveira in press).
This research, building on earlier work in the 1950s and
1960s (Almeida 1961a, 1961b; Almeida and Almeida 1959;
Almeida and Zbyszewski 1967; Glover 1970, 1986),
indicates that East Timor’s culture history follows patterns
found elsewhere in Island Southeast Asia. The earliest sites
dating to 30–40,000 BP are in limestone caves, probably
because caves provide better conditions for preservation and
site visibility. Between 4000 and 3500 years ago, there was
a shift from a hunting/fishing/gathering economy to a
farming economy, characterized archaeologically by the
presence of domesticated animals including pigs, chickens,
dogs, agriculture and earthenware pottery (Bellwood 1997,
2004). Goats were likely to have been introduced after
1000 AD. Although Glover claims a date of 5000 BP for
goats in East Timor, this is not supported anywhere else in
Island Southeast Asia (Glover 1986). Rice agriculture
Figure 1. East Timor.
appears to have been practiced in the Philippines, but early
agriculture in southern Island Southeast Asia, including
Timor, is poorly understood (O’Connor 2006; Spriggs
2003). It is likely to have included the cultivation of fruit
and nut bearing trees, sago and root crops such as taro and
yams (Latinis 2000; Stark and Latinis 1992). Land use
probably combined swiddening and utilization of existing
wetlands for appropriate root crops. The introduction of
metal tools to the region sometime after 2000 BP probably
allowed more intensive land clearance in a wider variety of
forest ecosystems than stone tools and fire alone. Irrigation
systems were not introduced into the region until the first
millennium AD, probably later than that in East Timor,
although there are few hard data on the history of agricultural practices.
Regarding settlement patterns, evidence from cave and
rockshelter sites in East Timor indicates continued use of
these areas from the earliest human presence on the island to
today. Caves are still used for ceremonial purposes and in
recent memory were places of refuge during warfare
(Pannell and O’Connor 2005). However, we have much less
evidence for human settlement outside of caves and
rockshelters in Island Southeast Asia. In East Timor, coastal
shell middens have been dated to the post 4000 BP era, and
we have found undated lithic scatters and the fortified sites
discussed in more detail below. In other parts of Island
Southeast Asia, a similar bias to cave sites exists for the
early “Neolithic” period (4000–2000 BP), with a few
exceptions (Lape 2000a; Simanjuntak and Forestier 2005).
More non-cave sites have been excavated dating to the post
2000 BP period, and these sites provide some insight into
settlement patterns. By this time, if not earlier, people were
living in houses clustered into villages and towns in much of
Island Southeast Asia, particularly the better documented
coastal areas (Bacus 1999; Bulbeck, et al. 2001; Bulbeck
and Prasetyo 2000). A rich documentary record, particularly
for times after the arrival of first sustained contact with
Europeans after 1512 AD, also supports the archaeological
evidence (e.g. Reid 1988, 1993; Stoler 1989; Wolters 1999).
Both lines of evidence also suggest that raiding and trading
were dominant in the social systems of Island Southeast
Asia, including East Timor (Fox and Soares 2000; Gunn
1999; Junker 1999).
Since 2003, we have been investigating the stone walled
structures common on certain parts of the landscape of the
Lautem and Manatuto districts on the north coast of East
Timor (Chao in press; Lape 2006). These sites appeared
similar to fortified settlements found in certain other parts of
Island Southeast Asia and Oceania, most of which date to
the same period, 1100–1700 AD (Field and Lape in
preparation). Based on an admittedly cave-biased
archaeological record, these fortified settlements represent a
dramatic shift in landscape use by people in north coastal
East Timor. Although it isn’t clear where people were living
outside of caves and rockshelters prior to 1100 AD, they
were not living on cliff edges or hilltops. Below, we present
our model of settlement locale choice that may explain
this shift.
Climate sensitive settlement choice model
Frameworks based on evolutionary ecology are useful for
incorporating environmental changes into models of human
behavior (e.g. Dyson-Hudson and Smith 1978; Kennett, et
al. 2006). These frameworks can be used to map the highly
complex adaptations to environmental changes by both
human and non-human organisms within ecosystems. An
externally introduced change, such as a rainfall anomaly,
will stimulate a variety of responses by organisms within
ecosystems, which in turn will stimulate a variety of
counter-responses and adaptations in linked organisms. By
definition, all organisms will adapt to new conditions (or
perish), and changing conditions provide both new
opportunities and challenges for these organisms. Human
cultural responses are of course extraordinarily complex, as
culture and memory cannot be reduced to biological
adaptation. However, it is a realistic project to model human
response options within an evolving ecological system,
incorporating factors such as predictability, risk, and bethedging (e.g. other papers in this issue, Kennett and Clifford
2004; Kennett and Kennett 2000).
Our model proposes a causal link between fortification
building and spatial and temporal changes in the availability
of rainfall-dependant resources. Although many factors
undoubtedly affected the choices available to people in East
Timor with regard to settlement locale and fortification
building, we assume that resource predictability was a
primary factor affecting these choices. Furthermore,
precipitation is a major limiting factor for agricultural
resources in contemporary north coastal East Timor and was
probably a major factor in the past. Late 20th century ENSO
warm phase events have been associated with drought in
most of Island Southeast Asia including Timor, and data
suggests that this held true at least for the late Holocene as
well. The model we present here tests the assumption
whether resource unpredictability in certain temporal and
spatial arrangements is linked to fortification building. This
test applies only to the initial construction of fortifications
on a landscape. We assume that once fortifications are built
and available to inhabitants, the nature of subsequent
conflicts and responses to environmental change will be
qualitatively different (M.W. Allen 2006; Ferguson 2006).
Our simple agent-based model has three primary
components which predict initial fortification building in
time and space:
1. People will choose to invest in building fortified
settlements only when they rely on spatially fixed and
restricted resources. Most archaeologically documented
occurrences of fortifications occur in agricultural
societies where land tenure is key to subsistence
(Ferguson 2006), or in non-agricultural societies where
key resources are dense and spatially restricted, such as
for salmon in Coastal Northwest North America
(Maschner and Reedy-Maschner 1998).
2. People will choose to invest in building fortified
settlements during periods of unpredictable resource
availability. “Predictability” is a moving target, as
changing conditions constantly influence memory and
therefore predictions of future conditions. Since our
understanding of climatic change remains temporally
coarse, we can simply say that if droughts start to occur
much more frequently, and when this increasing
frequency would be noticeable to people (significant
changes at century level periodicity), than we can assume
that memories of climate will not coincide with actuallyoccurring climate and are thus unpredictable.
3. Fortifications will appear in regions that a) contain
resources tolerant of climate change (creating conditions
for relative surplus) AND b) are adjacent to regions with
resources intolerant of climate change (creating
conditions for relative deficit). We predict that spatial
resource imbalance is a necessary requirement for initial
raiding and response to raiding (including but not limited
to fortification building). In these situations, people from
resource deficit areas will raid nearby regions with
surpluses, either to gain access to stored food, or to take
possession of resource-rich land. Once this process
begins, it may continue for a variety of reasons even if
resources become balanced again.
To test this model, we will map rainfall dependant
resource availability over time and space in East Timor, then
test whether fortifications preferentially appear in times and
places predicted by the model.
Food production, water and storage
For about 3500 years, people in East Timor have relied at
least in part for food on agriculture and domestic animals, as
discussed above. Wild plant and animal resources from both
terrestrial and marine contexts were also likely important
components of subsistence, and remain so today in many
parts of East Timor. Most forms of agriculture in East Timor
would have been highly dependant on sufficient and timely
precipitation, and wild plant and animal resources would
have also been dependant on rain, though perhaps less
critically. We emphasize that it is not simply precipitation
amount that affects resource availability, but also (and
perhaps more importantly) precipitation timing and stability.
In 2005, for example, East Timor experienced a severe food
shortage and starvation became a major concern to the UN
and other relief agencies. Manatuto, one of East Timor’s rice
production centers, actually produced an average amount of
rice on an annual basis that year. However, according to
local informants, the harvest season was delayed because of
the delay in the onset of the previous rainy season. Even
with the aid of complex irrigation systems, delayed rainfall
resulted in a severe food shortage because people tend to not
have enough storage to cope with an unexpectedly late
harvest. More serious and long-lasting droughts may result
in near total crop failure, for which only international aid
can save people from starving.
Precipitation anomalies can have a negative impact on
resources whether wetter or drier. For example, in dry years,
people may experience problems with swidden scheduling,
such as shortened growing seasons, germination failure
(desiccation) and increased predation from parasites and
animals (Conklin 1961, 1975). In extreme conditions there
may be drinking water shortages. Wet years can also
negatively affect resources in normally drier areas,
including problems with swidden scheduling, germination
failure (fungi, erosion), top soil erosion and inadequate
burning for weed control. Mosquito-borne diseases such as
malaria and dengue fever also tend to increase in wet years.
As demonstrated by recent El Niño effects in East Timor,
the immediate impacts of droughts would have been
disastrous for agricultural productivity, with widespread
crop failure and food shortages, exacerbated by a lack of
stored food. In our model, we predict that these droughts
would have been most acute if El Niños were widely spaced
in time, or were the first in a series of closely-spaced events.
Otherwise people would have adapted to predictable
drought by increasing food storage capacity (including the
use of domestic animals as storage units), developing trade
relationships, moving cultivation to higher rainfall zones
and/or selecting domestic plants and animals with greater
drought tolerance. As El Niño does not appear to have
significant impact on marine resources, we might also see a
shift to those resources as a buffer. These adaptations should
be visible archaeologically, and we plan to investigate these
in the future.
Drought effects food production, but it can also impact
drinking water supplies. Water is in short supply in the
limestone karst area of Tutuala, in the Lautem district.
Current residents of Tutuala village must walk several
kilometers down to a reliable water supply, for example.
Water sources are few and well-known, and underground
sources that can be tapped with wells are rare. We mapped
those in the Tutuala area to see if the fortified sites were
associated with them, but there was no clear association. In
the oral traditions, water sources were rarely “owned” by
one clan, but usually have shared ownership, suggesting
cooperative use of limited water. The stories also often
associate with stories of immigration from the low
limestone islands to the north like Leti.
In the extremely dry north coast of East Timor, it is only
viable to grow padi or wet rice in areas that are close enough
to permanent rivers to use irrigation. Rivers used for
irrigation therefore distort the effective precipitation
distribution by making water captured from the wet
highlands available in the dry lowlands, but only in areas
quite close to the river basin. The lower Laclo River valley,
which empties into the sea near Manatuto, is one such area.
Manatuto, while receiving less average annual precipitation
than Tutuala, has a more reliable water supply. The Laclo
River flows permanently, and even in the driest years when
flow ceases in some of its tributaries, the geological
formations in Manatuto are amenable to accessing
subsurface aquifers. This distortion effect is complicated. If
mapped, we would place the “effective” rainfall clines very
close together along the river edges determined largely by
local topographic relief, which affects people’s abilities to
construct irrigation channels and irrigate agricultural fields.
Therefore, the overall effect of negative rainfall anomalies
in this riverine system would be to increase the desirability
of lands close enough to the river to be useful for irrigated
gardening. Land even just a few hundred meters away from
the river may be totally unproductive if it is too difficult to
Precipitation pattern and seasonality in East Timor
Average annual precipitation is not evenly distributed across
East Timor. Rainfall varies by primarily by elevation, with
higher elevations generally receiving more precipitation
(Figure 2). Monsoon patterns also affect the distribution of
rainfall, causing the south to be generally wetter than the
north (Durand 2002). This uneven distribution of precipitation is the major cause (along with temperature, geomorphology, soil chemistry and other factors) of East Timor’s
ecologically diverse landscape. In some areas, precipitation
clines are closely spaced, with different niches in close
proximity to each other. The eastern part of the Lautem
district is one such area. While short term droughts would
affect agricultural viability in every cline, humans may have
been able to quickly adapt to lower rainfall by shifting
cultivation to wetter areas to the south or in higher
elevations by moving relatively short distances. In other
words, the same agricultural methods would still work by
moving to nearby areas, “following the rain” up or down
slope. Wild plant and animal resources would also
presumably been available in different zones depending on
average rainfall, and animals may have also simply moved
to different areas to find habitats for which they were best
adapted. In general, the proximity of several different
rainfall zones in eastern Lautem would have made it an
attractive place for farmers during unpredictable climates
compared with ecologically homogenous landscapes such as
the island of Leti to the north, or areas on the north coast to
the west of Lautem. In these areas, rainfall clines are more
widely spaced, which means that people in search of wetter
areas in times of drought would have to travel further. The
ecological diversity of these areas of closely spaced rainfall
clines may have also attracted immigrants, raiders and
traders looking for more productive agricultural land or
agricultural products.
Past ENSO and drought in East Timor
In recent years, data about late Holocene climate changes in
the Pacific have become available in more geographical
areas and at an increasing variety of time scales (e.g.
Moseley 1997; Moy, et al. 2002a; Salinger, et al. 1995).
While this period in Island Southeast Asia has still seen
relatively little attention from palaeoclimatologists,
researchers are beginning to outline broad climate patterns
for the region (Haberle 2000; Haberle, et al. 2001; Hope
2001). For the post 1000 AD era, ENSO-related climate
patterns appear to be a dominant environmental force in
Island Southeast Asia (Gagan, et al. 2004).
ENSO is a highly complex phenomenon with global
effects. In the recent past for which instrument data are
Figure 2. Precipitation in Dili, average vs. 1997-98 El Niño, showing lower average precipitation during dry season and
delayed onset of rainy season, adapted from Ministry of Forestry, Agriculture and Fisheries, Timor Leste
available, ENSO warm phases, or El Niños, are associated
with decreased rainfall on the order of 40–60% and slightly
decreased Sea Surface Temperatures (SST) in much of
Island Southeast Asia (M.S. Allen 2006; Field 2005; Folland
2002; Gagan et al. 2004). For example, the strong El Niño
of 1997–98 caused major droughts throughout much of
Island Southeast Asia, which exacerbated anthropogenic
burning to cause widespread, uncontrolled forest fires
(Terry, et al. 2001). ENSO induced SST variation is
relatively low in Island Southeast Asia due to the
moderating influence of the Indo-Pacific Warm Pool
(IPWP). Strong El Niños during the recent past have been
associated with relatively minor SST anomalies in the range
of 0.5 degrees in Island Southeast Asia. This suggests that,
unlike in west coast of the Americas, ENSO in Island
Southeast Asia is primarily a terrestrial phenomenon, and its
impact on marine ecosystems comes from erosion and
runoff effects in nearshore ecosystems rather than SST
Preliminary data suggests that increased interaction with
the Pacific Intertropical Convergence Zone (ITCZ) in the
mid Holocene enhanced ENSO and established modern
periodicity at about 5000 BP (Gagan et al. 2004). However,
the long term history of ENSO in Island Southeast Asia is
still under investigation and we cannot assume that recent
patterns held true throughout the post 1000 AD period. In
this paper we rely on evidence for ENSO frequency and
intensity from other regions and assume straightforward and
stable teleconnections to local Island Southeast Asian
conditions and stable geographical patterning. With regard
to this geographical pattering, Island Southeast Asia appears
to in the center of a zone of negative precipitation anomalies
during recent El Niño events, and therefore may be less
susceptible to minor spatial shifts in the phenomenon (in
contrast to the central Pacific, where even more caution is
advised). With regard to chronology, high resolution records
are now available for South America that can pinpoint the
timing and relative strength of ENSO events in the
Holocene and earlier with some certainty (Moy, et al. 2002a,
2002b). These data show a dramatic increase in El Niño
events well above modern levels from 1100–1600 AD, with
a peak from 1300–1400 AD.
Paleoclimate data typically does not have season-level
resolution. However, because recent ENSO warm phase
events have proceeded in a predictable manner over the
course of a calendar year, we can predict different effects
depending on the seasonality of normal rainfall patterns. For
example, in East Timor which is affected by the Southeast
Asian monsoon, most rain falls in the months of
November–March, with very limited rain during other
periods (Durand 2002). People have adapted to this pattern
in farming practices, scheduling burning towards the end of
the dry season in September–November, and planting as
soon as the first rains begin, usually late November–
December. In contrast, other parts of Island Southeast Asia
have the heaviest rain during the June–August monsoon, or
in some cases they may have two equal rainy seasons per
year. El Niño events are strongest during December–
Figure 3. Climate zones in East Timor, adapted from
Durand (2002).
February, which coincides with what are normally the
highest rainfall months in Timor. Thus Timor will experience a greater impact on total annual rainfall during El Niño
events compared with places which get the bulk of their rain
in June–August, when El Niño impact on precipitation is
typically much weaker. The overall effect of El Niño in East
Timor is that the dry season is extended and rainy season
shortened (Figure 3).
Fox and Soares (2000) illustrated in detail the effects of
this change in precipitation amount and timing during the
1997–98 El Niño on farmers’ decision making in the eastern
Lesser Sunda islands, including East Timor. These islands
were more affected by the 1997–98 El Niño than other parts
of Indonesia because of the prevalence of single-crop corn
(maize) farmers who are dependant on rains in December
and January. Farmers in the coastal regions of the eastern
Lesser Sundas lost roughly 50% of their corn crop during
this El Niño (Fox and Soares 2000:181). The study found
that farmers had sufficient seed reserves for no more than
two attempts at planting, as increasing seed reserves came at
the expense of food reserves. Although corn did not exist in
East Timor and any other places in Southeast Asia before the
European contact in the early 16th century, the unpredictability of precipitation timing and stability during El Niño
events may have had similar catastrophic impacts on other
crops with similar scheduling regimes.
Fortified sites in the Lautem district
The stone structures near the villages of Tutuala and Ira Ara
in the Lautem district play an important role in
contemporary cultural practice as sacred places, sites of clan
histories and places of other kinds of social memory (Lape
2006). They are considered to be old village sites (lata irinu
in the regional Fataluku language) by local residents
(McWilliam 2002, 2003; Pannell and O’Connor 2005). The
remains of fortified settlements vary somewhat in form, but
generally they consist of dry-stacked limestone rock walls
from 1.5–4 m in height, and 1–3 m in width at their base
enclosing areas of 500–3000+ m2. The sites are located on
hilltops or cliff edges. In the latter case, walls tend not to
extend along the cliff edge suggesting that the steep drop off
served a defensive purpose. Stone features (e.g. platforms
and lower walls) are found enclosed within these outer
walls. The exterior walls have an opening or doorway, and
in some cases the entrance is walled for several metres to
form a narrow twisting hallway. Many of these sites also
contain sacred wood or stone markers (sikua or saka in
Fataluku). These markers, along with stone platforms (chaluluturu, Fataluku for “ancestral grave”), are a focus of contemporary ritual practice (see also McWilliam 1991, 2001,
2002, 2003; Pannell 2004; Pannell and O’Connor 2005).
We have identified most of the fortified sites existing on
the landscape in the surveyed areas marked in Figure 4.
These include eighteen fortified sites in the Tutuala area and
two in the Ira Ara area. Thousands of years of swidden
burning and farming in this steeply sloping region, as well
as occupation activities themselves, have contributed to
sediment disturbance and erosion. This has had implications
for dating occupation periods (discussed in detail in Lape
2006). A fortified site in the village of Ira Ara proved to have
deeper deposits with intact stratigraphy when excavated in
2005. Dates obtained for the Tutuala sites range from 2300
BC–1920 AD, although as discussed in Lape 2006, these
sites were likely fortified in the 15th–19th centuries AD.
The Ira Ara site is more securely dated to the 15th–19th
centuries AD (Table 1).
Pottery and faunal assemblages recovered from test
excavations are similar to those recovered from Neolithic
occupation sites in caves in the area, such as at Lene Hara
and in open sites, such as Tutunchau. These include
fragmented undecorated earthenware pottery, marine shell,
and animal bone. The presence of food remains and
utilitarian pottery in these assemblages suggest that these
sites were domestic spaces with long term occupation, and
not just military redoubts or short term refuges, although
this is subject to further testing and analysis. Oral traditions
in Tutuala and Ira Ara also support the view that these sites
were fortified villages which were continuously occupied
(until the population moved to the next place). People in
Tutuala and other parts of East Timor remember a time of
perang saudara or perang ratu (Indonesian), literally
“brotherly war” or “clan war”, which predated Portuguese
colonial administration. Headhunting and slave raiding
themes are also common in clan histories stories, echoing a
long tradition of headhunting in Southeast Asia (Hoskins
1996). Additionally, many place names, clan names and
histories suggest migration from other places. These names
(such as Leti, Oirata, Malei, and even China) can be traced
to nearby and distant islands. In some cases, familial and
economic ties were maintained with these places until quite
recently (Josselin de Jong 1937).
While documentary history about the Tutuala area is
restricted to the 20th century, other parts of Timor were
described by the first chroniclers as having a large number
of small polities, based in the interior mountains and
perpetually at war with each other (Pigafetta 1969; Pires and
Rodrigues 1944; Prapañca and Robson 1995; Ptak 1998;
Reid 1985, 1988; Wallace [1869] 1986), a feature of other
places in pre-colonial Island Southeast Asia as well (Junker
1999; Lape 2000b, 2000c). By the middle of the 20th
century, the use of most indigenous fortified villages had
changed significantly from places of permanent occupation
to irregular use for ceremonial activities and for strategic
military sites in the 25-year long resistance by East
Figure 4. Lautem area sites.
site name
depth below
surface (cm)
lab number
age (BP)
calibrated date*
Ira Ara
Ira Ara
burial 1
burial 1
marine shell
600 +/- 40
80 +/- 40
Ira Ara
Ira Ara
Lama #4
Bukit Aiteas
2004 TP1 L5
2005 TP03 F1
2005 TP03 L7
2005 TP02 L5
burial 2
burial 2
human bone
360 +/- 50
310 +/- 40
Bukit Aiteas
Bukit Aiteas
marine shell**
1310± 40 BP
780± 55 BP
AD 1650-1810
AD 1680-1740
AD 1800-1930
AD 1950-1960
AD 1440-1650
AD 1470-1660
AD 1460-1660
AD 1800-1950
AD 1452-1668
AD 649-878
AD 1420-1530
AD 1540-1640
AD 1430-1520
AD 1560-1630
AD 1015-1202
AD 1150-1300
Table 1. Radiocarbon dates of fortified sites (see Lape 2006 for additional dates for Lautem area sites).
*Calibration for radiocarbon dates to 2-sigma calculated using INTCAL 98 (Stuiver 1998)
**Terebralia sulcata, marine reservoir correction not applied
Timorese to Indonesian annexation (Pannell and O’Connor
in press).
Fortified sites in the Manatuto district
From 2004 to 2006, we conducted a survey and excavated a
series of test pits in the Manatuto area on the north coast of
East Timor. The survey area, about 40 km2, covers the lower
floodplain of the Laclo River and the Lamessana lagoon to
the east (Figure 5). In general, the lowlands in this triangleshape area are flooded annually, and it is bordered by at least
two steps of marine terraces and adjacent mountains up to
500–800 meters in elevation. We attempted full-coverage
Figure 5. Manatuto area sites. Fortified sites include: 1) Lekpaturen, 2) Soraha, 3) Soraha Barat,
4) Bukit Aiteas, 5) Malarahun Lama #1, 6) Malarahun Lama #4, 7) Hataro #5, 8) Luri.
Period I
AD 1400-1550
Period II
AD 1550-1650
Period III
AD 1650-1725
Period IV
AD 1725-1850
Period V
post AD 1850
Soraha Barat
Bukit Aiteas
Malarahun Lama #1
Malarahun Lama #4
Hataro #5
Total (identified sherds)
Table 2. Asian tradeware seriation from Manatuto area fortified sites, all samples from surface collections. Number (%).
survey except where there was no surface archaeological
visibility, such as in the rice paddies where we only
surveyed on the dikes and uncultivated edge areas near the
villages and foothills. In the less vegetated terraces, random
walkovers covered most of survey area. Our three-season
survey has identified 60 sites and at least 22 isolated finds.
Eight of these sites are located on hilltops with substantial
walling and terracing and have been dated to the 11th to
early 18th century AD based on radiocarbon dates and Asian
tradeware seriation (Tables 1 and 2) (Chao in press).
The eight fortified hilltop sites all have restricted access
and far-reaching views overlooking the Laclo River mouth
and lowland plains. The sizes of these sites vary from 100 to
200 metres in length and they have similar layouts. In the
center of the site, there often is an oval or rectangle-shaped
raised platform of varying heights, which has rock-piled
features on the corners. In some cases, these platforms are
the focus of contemporary ritual practice, such as sacrificial
rites before planting and after harvest. Surrounding the
platform, several levels of terraces are divided by lines of
low rock-piled walls and/or dry-stacked walls of 1–2 metres
in height. In most cases, walls extend to the cliff edge and
the steep drop-off or slopes in all directions would have
served a natural defensive function. Examples include the
Bukit Aiteas, Malarahu Lama #1 and #4 sites, which may
cluster into one large site connected through the path on the
hilltop ridge.
Four hilltop sites were excavated: Lekpaturen, Malarahun Lama #4, Bukit Aiteas and Soraha. Soraha lacked
stratigraphic integrity, so we did not date any samples. Other
sites had sufficient stratigraphic integrity and produced
good radiocarbon sequences, supported by stylistic dates on
imported tradeware. Radiocarbon dates on charcoal samples
from Lekpaturen and Malarahun Lama #4 indicate that these
sites were likely first occupied between 1450 AD and 1650
AD. The earlier date of 650–880 AD at Lekpaturen is
probably anomalous as the sample was recovered from a
layer containing post-1550 AD tradeware. Bukit Aiteas
dates indicate an initial occupation as early as 1150–1300
AD (Table 1). As with the Tutuala and Ira Ara dates, they
fall in an unfortunate area of the radiocarbon calibration
curve and return uncertainties that can span two centuries. In
addition, this precision is enhanced for the Manatuto area
sites by Asian tradeware ceramics found in pottery
assemblages (tradeware was not present in the Tutuala and
Ira Ara sites). On the basis of the 5-phase chronology from
surface collected tradeware samples (Table 2), the hilltop
sites were in Manatuto mostly occupied from mid 16th to
early 18th century AD, while as suggested by the excavated
radiocarbon samples, some sites would have been occupied
somewhat earlier, before tradeware was widely exchanged
in Island Southeast Asia. These dates fall within the
radiocarbon chronology and represent a refinement of that
chronology, with the normal caveats associated with
stylistic dating.
It should be noted that the survey area has been
continuously inhabited for at least three millennia, but these
hilltop sites are the only places that people chose to live on
by the 16th and 17th centuries AD, while the numerous
earlier sites that we have found on the terraces and
floodplain were abandoned during this period. Many of
these earlier sites also lie further from the Laclo river basin
(Figure 5). While some of these areas are now connected to
the extensive modern irrigation system, others (such as the
Lamessana Lagoon area) cannot be irrigated without
mechanical pumps.
Testing our model
Our model predicts that fortifications would appear on a
landscape under three conditions: 1) where resources are
spatially constrained; 2) when they are temporally
unpredictable; and 3) in relatively resource rich (buffered)
areas that are next to relatively resource poor areas. The
record of fortifications on the landscapes of Tutuala, Ira Ara
and on the lower Laclo River in Manatuto meets these
conditions. a) All three areas were largely dependant on
agriculture for subsistence. b) In all three areas,
fortifications appear on the landscape primarily between
1150–1550 AD. This corresponds closely to the 1100 to
1600 AD period when El Niño frequency increased above
earlier periods. However, the fortification building peak of
1450–1650 comes somewhat later than the El Niño event
frequency peak in 1300–1400 AD. c) Fortifications in these
three areas are situated in or near areas that would have had
relatively viable agriculture (through rainfall or irrigation)
even during extended droughts, and are adjacent to areas
that lack drought-tolerant agricultural potential.
This final condition requires some additional discussion.
In Tutuala, irrigation is impossible and past agriculture must
have been swidden, as it is today. While swiddening
assumes mobility (fields are currently fallowed after 3–5
years of use), this mobility is currently constrained by clanlevel land ownership systems that limits movement within
defined territories. We must assume that people in the past
would have also been limited in their ability to find new
viable agricultural lands far from their home base for our
data to fit our model. Those clans whose territories included
multiple rainfall zones or clines would have had an advantage over those who did not during periods of decreasing
and unpredictable rainfall. Jaco Island, which falls entirely
within one rainfall cline and has no groundwater, had only
one fortified site, a much lower density than that recorded in
the mainland Tutuala area. In the Ira Ara area, rainfall clines
are spaced further apart and fortification density is lower
than in Tutuala. However, the Ira Ara site itself is directly
adjacent to a spring (ira ara means “water source” in
Fataluku), which probably allowed for limited irrigation in
the slopes below to the sea in the past as it does now. In the
Manatuto area, pre-12th century sites are located across the
entire survey area. During later periods, when we see
evidence of decreasing and unpredictable rainfall, only
areas on hilltops very close to the Laclo River have sites,
and these sites are all fortified. We interpret this pattern as
evidence that people preferentially settled and fortified
hilltops directly adjacent to these viable areas during these
times, in defense against those who live in more distant
regions who were suffering food shortages.
Some of the data does not obviously fit the model. There
are several pre-1000 AD dates from fortified settlements in
both Manatuto (explained above) and Lautem (Lape 2006).
Given the over 35,000 year occupation history of East
Timor, it is possible that these early dates represent prefortification use of these hilltop or cliff edge sites. A more
problematic disjunction between the model and the data is
that the majority of the dates for the fortified sites cluster
around the 1450–1650 AD period, which is later than the El
Niño frequency peak of 1300–1400 AD. As suggested
above, these later sites may have been built as a result of
social forces indirectly related to resource shortages, and
may have instead been the adaptation to a system that had
already begun to be fortified several hundred years earlier.
The large radiocarbon uncertainty may also explain this
apparent discrepancy. Despite this possible temporal
disjunction between resources and fortifications, our spatial
data still suggest that fortifications were built preferentially
in areas containing drought resistant resources.
In summary, the evidence presented here suggests that
fortification building was associated with decreasing and
unpredictable rainfall in time, and in areas of drought
tolerance next to areas lacking tolerance in space. We plan
to further refine and test this model in the near future, both
in East Timor and other areas. The present case study would
be improved by a) refining the construction and occupation
chronology for the fortifications through additional survey
and excavation; b) further testing of the model by surveying
in other areas with a variety of real or effective precipitation
clinal densities; and c) collecting local palaeoclimatic proxy
records to supplement global records. Further improvements
would come from a better understanding of past agricultural
crops and practices and of the use of marine resources as
potential buffers during droughts in East Timor. We also
have only an outline of past regional interactions and the
frequency of trading and raiding in the region. While
currently lacking, all of this data should be obtainable in the
East Timor archaeological and paleoclimate record.
More broadly, this model could be applied to a variety of
situations globally. Cases where rainfall limited agricultural
production and where ENSO systems affected rainfall
would be good candidates for comparative study, such as the
western Pacific or the Colorado Plateau region of North
America. Areas with other types of spatially constrained
resources that might have significantly varied over time,
such as the Pacific Northwest of North American reliant on
salmon, might also be candidates for application of this
model. Ultimately, we hope the approach we have detailed
in this paper can be used to move beyond simple
deterministic interpretations of past landscape use and begin
to resolve the complex factors that influence human social
Research in East Timor conducted with the permission and
cooperation of the Ministry of Education and Culture and
was funded by the University of Washington and Academia
Sinica. The authors thank Cecilia Assis and her staff at the
Ministry of Culture, the people of Tutuala, Ira Ara and
Manatuto and the 2005 University of Washington field
school students for their assistance, cooperation and support
of the fieldwork. Thanks to Douglas Kennett, Peter White
and an anonymous reviewer for helpful comments that
improved this paper. We take responsibility for remaining
Allen, M.S. 2006 New ideas about late Holocene climate
variability in the Central Pacific. Current Anthropology
Allen, M.W. 2006 Transformation in Maori warfare: Toa, Pa
and Pu. In The Archaeology of Warfare, edited by E. Arkush
and M. W. Allen, pp. 184-213. University Press of Florida,
Almeida, A.D. 1961a Contribution à l’étude du néolithique du
Timor Portugais. Paper presented at the 5th International
Congress of Prehistoric and Protohistoric Sciences,
––––1961b Presenças culturais no Timor Portugues. Estudos
de Ciencias Politicas e Sociais. Copies available from 51.
Almeida, A.D. and M.E.D.C.E. Almeida 1959 Contribuiçao
para o estudo seroantropologico dos Macaenses. Estudos
sobre Antropologia Fisica do Ultramar Portugues 1(13):
Almeida, A.D. and G. Zbyszewski 1967 Contribution to the
study of the prehistory of Portuguese Timor – lithic industries. Asian and Pacific Archaeology Series (1):55-67.
Bacus, E. 1999 Prestige and potency: Political economies of
protohistoric Visayan polities. In Complex Polities in the
Ancient Tropical World, edited by E. Bacus and L. Lucero,
pp. 67-87. American Anthropological Association,
Bellwood, P. 1997 Prehistory of the Indo-Malaysian
Archipelago. Rev. ed. University of Hawaii Press, Honolulu.
2004 The origins and dispersals of agricultural communities in
Southeast Asia. In Southeast Asia from Prehistory to
History, edited by I. Glover and P. Bellwood, pp. 21-40.
RoutledgeCurzon, New York.
Bulbeck, D., M. Pasqua and A.D. Lello 2001 Culture History
of the Toalean of South Sulawesi, Indonesia. Asian
Perspectives 39(1/2):79-108.
Bulbeck, F.D. and B. Prasetyo 2000 Two millenia of sociocultural development in Luwu, South Sulawesi, Indonesia.
World Archaeology 32(1):121-137.
Chao, C.-Y. [in press] Tsu-taing Tong-ti-wen Ku-tai Mau-i
Kao-ku (A preliminary archaeological investigation toward
ancient trade in Manatuto, East Timor). Tieng-yieh Kao-ku
(Field Archaeology of Taiwan) 10.
Conklin, H.C. 1961 The study of shifting cultivation. Current
Anthropology 2(1):27-61.
––––1975 Hanunoo agriculture: a report on an integral system
of shifting cultivation in the Philippines. Elliot’s Books,
Northford, Conn.
Durand, F. 2002 Timor Lorosa’e, pays au carrefour de l’Asie et
du Pacifique: un atlas gâeo-historique. Presses universitaires de Marne-le-Vallâee; IRASEC. Institut de recherche
sur l’Asie du sud-est contemporaine.
Dyson-Hudson, R. and E.A. Smith 1978 Human Territoriality:
An Ecological Reassessment. American Anthropologist
Ferguson, R.B. 2006 Archaeology, anthropology and the
origins and intensification of war. In The Archaeology of
Warfare, edited by E. Arkush and M.W. Allen, pp. 469-524.
University Press of Florida, Gainesville.
Field, J. and P. Lape in preparation Climatic variance and the
emergence of fortifications in Island Southeast Asia and
Field, J.S. 2005 Interannual rainfall, ENSO and agricultural
productivity in the Western Pacific. In The Renaca Papers:
VI International Conference on Rapa Nui and the Pacific,
edited by C.M. Stevenson, J.M. Ramirez, F.J. Morin and N.
Barbacci, pp. 75-80. Easter Island Foundation, Los Osos.
Folland, C. 2002 Relative influences of the interdecadal Pacific
Oscillation and ENSO on the South Pacific Convergence
Zone. Geophysical Research Letters 29(13):21-1 21-4.
Fox, J.J. and D.B. Soares (editors) 2000 Out of the Ashes:
Destruction and reconstruction of East Timor. Crawford
House, Adelaide.
Gagan, M., Hendy, E., Haberle, S., Hantoro, W. 2004 Postglacial evolution of the Indo-Pacific warm pool and El
Niño-Southern Oscillation. Quaternary International 118119:127-143.
Glover, I. 1970 Excavations in Timor: a study of economic
change and cultural continuity in prehistory. PhD dissertation, Australian National University.
1986 Archaeology in Eastern Timor, 1966-67. Terra Australis
11. Research School of Pacific Studies, Australian National
University, Canberra.
Gunn, G. 1999 Timor Loro Sae: 500 years. Livros do Oriente,
Haberle, S.G. 2000 Vegetation Response To Climate
Variability: A Palaeoecological Perspective on the ENSO
Phenomenon. In El Niño: History and Crisis, edited by R.
Grove and J. Chappell, pp. 66-78. White Horse Press,
Haberle, S.G., G.S. Hope and W.A. Van der Kaars 2001
Biomass burning in Indonesia and Papua New Guinea:
natural and human induced fire events in the fossil record.
Journal of Palaeogeography, Palaeoclimatology and
Palaeoecology 171:259-268.
Hope, G.S. 2001 Environmental change in the Late Pleistocene
and later Holocene at Wanda Site, Soroako, South Sulawesi,
Indonesia. Journal of Palaeogeography, Palaeoclimatology
and Palaeoecology 171:129-145.
Hoskins, J. (editor) 1996 Headhunting and the Social
Imagination in Southeast Asia. Stanford University Press,
Hughen, K.A., M.G.L. Baillie, E. Bard, J.W. Beck, C.J.H.
Bertrand, P.G. Blackwell, C.E. Buck, G.S. Burr, K.B. Cutler,
P.E. Damon, R.L. Edwards, R.G. Fairbanks, M. Friedrich,
T.P. Guilderson, B. Kromer, G. McCormac, S.W. Manning,
C.B. Ramsey, P.J. Reimer, R.W. Reimer, S. Remmele, J.R.
Southon, M. Stuiver, S. Talamo, F.W. Taylor, J.v.d. Plicht
and C.E. Weyhenmeyer 2004 Marine04 marine radiocarbon
age calibration, 0-26 Cal KYR BP. Radiocarbon
Josselin de Jong, J.P.B.d. 1937 Oirata: A Timorese settlement of
Kisar, Studies in Indonesian Culture. Verhandelingen der
Koninklijke Nederlandse Akademie van Wetenschappen,
Afdeeling Letterkunde Nieuwe Reeks Deel XXXIX.
Junker, L. 1999 Raiding, Trading, and Feasting: The political
economy of Philippine chiefdoms. University of Hawai`i
Press, Honolulu.
Kennett, D.J., A. Anderson and B. Winterhalder 2006 The Ideal
Free Distribution, Food Production, and the Colonization of
Oceania. In Behavioral Ecology and the Transition to
Agriculture, edited by D.J. Kennett and B. Winterhalder.
University of California Press, Berkeley.
Kennett, D.J. and R. Clifford 2004 Flexible Strategies for
Resource Defense on the Northern Channel Islands of
California: An Agent-Based Model. In The Archaeology of
Insularity: Examining the Past in Island Environments,
edited by S. Fitzpatrick, pp. 21-50. Greenwood Press,
Westport, CT.
Kennett, D. J. and J. P. Kennett 2000 Competitive and
Cooperative Responses to Climatic Instability in Southern
California. American Antiquity 65(2):379-395.
Lape, P.V. 2000a Contact and Conflict in the Banda Islands,
Eastern Indonesia, 11th–17th Centuries. PhD Dissertation,
Brown University.
––––2000b Contact and Conflict in the Banda Islands, Eastern
Indonesia, 11th–17th Centuries. PhD Dissertation, Brown
––––2000c Political dynamics and religious change in the late
pre-colonial Banda Islands, Eastern Indonesia. World
Archaeology 32(1):138-155.
––––2006 Chronology of Fortified Settlements in East Timor.
Journal of Island and Coastal Archaeology 1(2):285-296.
Lape, P.V., S. O’Connor and N. Burningham in press Rock art:
A potential source of information about past maritime
technology in the South-East Asia-Pacific region.
International Journal of Nautical Archaeology.
Latinis, K. 2000 The development of subsistence system
models for Island Southeast Asia and Near Oceania: The
nature and role of arboriculture and arboreal-based
economies. World Archaeology 32(1) 41-67.
Maschner, H.D.G. and K.L. Reedy-Maschner 1998 Raid,
Retreat, Defend (Repeat): The Archaeology and Ethnohistory of Warfare on the North Pacific Rim. Journal of
Anthropological Archaeology 17:19–51.
McWilliam, A. 1991 Prayers of the sacred stone and tree:
aspects of invocation in West Timor. Canberra Anthropology 14(2):49-59.
––––2001 Prospects for the sacred grove: Valuing lulic forests
on Timor. The Asia Pacific Journal of Anthropology 2(2):89113.
––––2002 Timorese seascapes: Perspectives on customary
marine tenure in East Timor. The Asia Pacific Journal of
Anthropology 3(2):6-32.
––––2003 New beginnings in East Timorese forest
management. Journal of Southeast Asian Studies 34(2):307327.
Moseley, M.E. 1997 Climate, culture, and punctuated change:
new data, new challenges. Review of Archaeology 17:19–27.
Moy, C.M., G.O. Seltzer, D.T. Rodbell and D.M. Anderson
2002a Variability of El Nino/Southern Oscillation activity at
millennial timescales during the Holocene epoch. Nature
––––2002b Laguna Pollcacocha Sediment Color Intensity
Data IGBP PAGES. World Data Center for Paleoclimatology Data Contribution Series #2002-76. NOAA/NCDC
Paleoclimatology Program, Boulder.
O’Connor, S. 2003 Nine new painted rock art sites from East
Timor in the context of the Western Pacific region. Asian
Perspectives 42(1):96-128.
2006 Unpacking the Island Southeast Asian Neolithic Cultural
Package, and Finding Local Complexity. In New
Perspectives on Early Southeast Asia: Proceedings of the
10th International Conference of the European Association
of Southeast Asian Archaeologists, London 14th–17th
September 2004, edited by L. Bacus and I. Glover.
Singapore University Press, Singapore.
O’Connor, S. and N.V. Oliveira in press Inter and Intra
Regional Variation in the Austronesian Painting Tradition: A
View from East Timor. Asian Perspectives.
O’Connor, S. and P. Veth 2005 Early Holocene shell fish hooks
from Lene Hara Cave, East Timor establish complex fishing
technology was in use in Island South East Asia five
thousand years before Austronesian settlement. Antiquity
O’Connor, S., M. Spriggs and P. Veth 2002 Excavation at Lene
Hara Cave establishes occupation in East Timor at least
30,000–35,000 years ago. Antiquity 76:45–50.
Pannell, S. 2004 Welcome to The Hotel Tutuala: Fataluku
Accounts of Going Places in an Immobile World. Paper
presented at the Australian Anthropological Society
Conference, University of Melbourne.
Pannell, S. and S. O’Connor 2005 Toward a Cultural
Topography of Cave Use in East Timor: A Preliminary
Study. Asian Perspectives 44(1):193-206.
––––[in press] Where the wild things are: An exploration of
sacrality, danger and violence in confined spaces.
University of Colorado Press, Boulder.
Pigafetta, A. 1969 The voyage of Magellan. Translated by P.S.
Paige. Prentice-Hall, New Jersey.
Pires, T. and F. Rodrigues 1944 The Suma Oriental of Tomé
Pires, An account of the East, from the Red Sea to Japan,
written in Malacca and India in 1512-1515, and The book of
Francisco Rodrigues, rutter of a voyage in the Red Sea,
nautical rules, almanack and maps, written and drawn in
the East before 1515. Translated by A. Cortesão. Ser. 2, no.
89-90. 2 vols. The Hakluyt Society, London.
Prapañca and S.O. Robson 1995 Desawarnana: (Nagarakrtagama). Verhandelingen van het Koninklijk Instituut
voor Taal-, Land- en Volkenkunde; 169. KITLV, Leiden.
Ptak, R. 1998 Glosses on Wang Dayuan’s Daoyi Zhilue. In
China and the Asian Seas: Trade, travel, and visions of the
other (1400-1750), edited by R. Ptak, pp. 127-141. Ashgate,
Aldershot; Brookfield, Vt.
Reid, A. 1985 Trade goods and trade routes in Southeast Asia:
C.1300-1700. SEAMEO Project in Archaeology and Fine
––––1988 Southeast Asia in the Age of Commerce, 1450-1680
Volume One: The Lands Below the Winds. Yale University
Press, New Haven.
––––1993 Southeast Asia in the Age of Commerce, 1450-1680
Volume Two: Expansion and Crisis. 2 vols. Yale University
Press, New Haven.
Salinger, M.J., R. Basher, B. Fitzharris, J. Hay, P. Jones, J.
McVeigh and I. Schmidely-Leleu 1995 Climate trends in the
south-west Pacific. International Journal of Climatology
Simanjuntak, T. and H. Forestier 2005 Research Progress on
the Neolithic in Indonesia. In Southeast Asian Archaeology:
Wilhelm G. Solheim II Festschrift, edited by V. Paz, pp. 84104. University of the Philippines, Quezon City.
Spriggs, M. 2003 Chronology of the Neolithic transition in
Island Southeast Asia and the Western Pacific: a view from
2003. Review of Archaeology 24(2):57-80.
Spriggs, M., S. O’Connor and P. Veth 2003 Vestiges of early
pre-agricultural economy in the landscape of East Timor:
Recent research. In Fishbones and Glittering Emblems:
Proceedings from the EurASEEA Sigtuna Conference, edited
by A. Kallen and A. Karlstrom, pp. 49-58. Museum of Far
Eastern Antiquities, Stockholm.
Stark, K. and K. Latinis 1992 The archaeology of sago
economies in Central Maluku: An initial sketch. Cakalele
Stoler, A.1989 Rethinking colonial categories: European
communities and the boundaries of rule. Comparative
Studies in Society and History 31:134-161.
Stuiver, M., P.J. Reimer, E. Bard, J., W. Beck, G.S. Burr, K.A.
Highen, B. Kromer, G. McCormac, J. van der Plicht and M.
Spurk 1998 INTCAL98 Radiocarbon Age Calibration,
24000-0 cal BP. Radiocarbon 40 (3) 1041-83.
Terry, J.P., R. Raj and R.A. Kostaschuk 2001 Links between the
southern oscillation index and hydrological hazards on a
tropical Pacific island. Pacific Science 55:275-283.
Wallace, A.R. [1869] 1986 The Malay Archipelago: The land of
the orang-utan, and the bird of paradise. Oxford University
Press, Oxford.
Wolters, O.W. 1999 History, Culture, and Region in Southeast
Asian Perspectives. Southeast Asia Program Publications,
Cornell University, Ithaca, NY.